Common Dispersion in EdgeR

[bob-loblaw]

Hi all,

I had a question about the common dispersion in EdgeR. In the manual it gives the following..

> sqrt(d$common.disp)
[1] 0.416

The square root of the common dispersion gives the coefficient of variation of biological variation (BCV). Here the BCV is 41%

However in my own dataset...
> sqrt(y$common.disp)
[1] 2.096939


So is the BCV for my dataset 200%? Is this even possible?

Edit: Am I correct in saying that this would mean the true abundance for each gene can vary by up to 200% between samples?

I want to be able to use this for the RNASeqPower bioconductor package. Now the version of EdgeR I'm using is 3.8.2 and the manual is for 3.7.17 on the bioconductor site, maybe there is a difference between versions.. but this is the manual that the package calls.

Thanks.